The Canvas AI Agents with polling will show the end-user a plan of the next step and inform the user that a task has been completed so they know whats happening, know to ask AI to stop what its doing if its not useful and don't feel as bad waiting too long for an answer as they know AI is doing something.
However, AI does regularly have some kind of plan for all the steps required to complete the task but this is difficult to obtain reliably and can change. We have discussed here how to do a true plan here: https://www.drupal.org/project/canvas/issues/3531000#comment-16241366 📌 Integrate incremental agent loop execution in XB AI Active
This is an issue once the previous issue is completed to improve things and output a full plan as per the designs.
Active
1.0
AI
Not all content is available!
It's likely this issue predates Contrib.social: some issue and comment data are missing.
No activities found.